work that is not the work

2022-10-10

attempting to make a profit is hard.

> TODO: insert rule of acquisition here

to save money, i buy used and commodity hardware.

now i spend my time figuring out server rails and how to get rid of fiber channel.

when i'd like to be spending my time debugging interactions between nomad consul and vault

which is still not my main purpose which is to

- develop software services and platforms for regional businesses and individuals

- create games and websites which rely heavily on backend servers

- provide limited web services which can scale

- create utility for professional software developers,

- especially if it will greatly aid fledgling beginners.

and that's... ok. but not profitable. even still, it's frustratingly slow and isn't easy to speed up. at least when writing software, there is opportunity to build/fail fast, and even to get things out of the door in an alpha state. but this is all backend, private cloud nonsense. basically, i don't trust VPS with my sensitive data, and i don't want to risk accidents from the nightmare land of becoming popular and then getting a 10k bill for 3 weeks of AWS. and that's like, a 'nice' bad day in AWS land. no thank you. i like vultr because it's prepay and reasonably priced, but that's basically a fancy vps it's certainly no AWS or GCP. eventually i'll be multicloud, once the private cloud is operational, once my iac is locked down, once i have at least one or two apps running fault tolerant in _my cloud_, then i'll give aws money.

in the mean time, i'll just give amazon money for all the things i can't get on fs.com or ebay or server rack online or wherever.

i've also now got a bunch of other little projects that are 'software' and needed in the long run but.. aren't what i'd theoretically 'like' to be doing.

like i have to figure out that onie open router / sonic stuff now because i've got an arista qsfp switch. and i think i need to switch to openwrt in stead of pfsense for my pcengines.ch apu boards. pcengines i like because i enjoy their power efficiency and they fill a nice niche. limited to about a gigabit, but the main purpose will be simple routing, nat, dhcp. the issue is that pfsense doesn't quite run at full network speed even without wireguard. i don't plan to use wireguard yet on these, but it's the easiest example to explain the issue. this blog post was the final nail for me

https://teklager.se/en/knowledge-base/apu2c0-ipfire-throughput-test-much-faster-pfsense/

this switch is annoying in the short term, but long-term as time passes and maybe i upgrade other equipment, having been on openwrt and experienced will allow an easy switch to installing 4g or 5g radios and allowing the pcengines boards to transition to edge networking. buuut i have not any money for the chips, its better to wait anyways, and i'm not buying however many sims and phone plans to try and execute on that. no need. very busy with the single internet connection i already have. but yeah, the apu boards i bought are used specific for telecom/light industrial in stead or addition to wifi. i wouldn't use an apu as a wifi AP, but i could see a backup phone sim with 35gb of data or something.

eventually i'd like to get more written down, maybe make some youtube videos and shorts, maybe do some live video. there is possibly room for more tech content on youtube.